Factors affecting the absorption maxima of acidic forms of bacteriorhodopsin

Transcription

1 E * S :* 0l Factors affecting the absorption maxima of acidic forms of bacteriorhodopsin A study with artificial pigments A. Albeck,* N. Friedman,* M. Sheves,* and M. Ottolenghi4 *Department of Organic Chemistry, the Weizmann Institute of Science, Rehovot 76100; and *Department of Physical Chemistry, The Hebrew University of Jerusalem, Jerusalem 91904, Israel ABSTRACT The absorption maximum (568 nm) of light-adapted bacteriorhodopsin br568 undergoes reversible changes after acidification. At ph 2.9, the absorption shifts to 605 nm (forming br605) and it blue shifts to 565 nm, after further acidification to ph (forming br565). Molecular models accounting for such acid-induced changes are relevant to the structure and function of bacteriorhodopsin. In the present study we approached the problem by applying artificial br pigments based on selectively modified synthetic retinals. This may allow direct identification of the specific regions in the retinal binding site where the above changes in the protein-retinal interactions take place. We investigated the spectroscopic effects of acid in a variety of artificial pigments, including cyaninelike retinals, retinals bearing bulky groups at C4, short polyenes, and retinals in which the f3-ionone ring was substituted by aromatic rings. The results provide direct evidence for the hypothesis that the generation of br6m is due to changes in polyene-opsin interactions in the vicinity of the Schiff base linkage. The second transition (to br%5) was not observed in artificial pigments bearing major changes in the ring structure of the retinal. Two approaches accounting for this observation are presented. One argues that the generation of brn5 is associated with acid-induced changes in retinalprotein interactions in the vicinity of the retinal ring. The second involves changes in polyene-opsin interactions in the vicinity of the Schiff base linkage. For both brw and brn5, our results do not discriminate between the direct titration of negative or dipolar protein groups in the binding site and changes in the retinal-protein interactions induced by changes in the protein structure outside of the binding site. INTRODUCTION Light-adapted bacteriorhodopsin (br568), the all-trans retinal-protein complex in the purple membrane of Halobacterium halobium, functions as a light-driven proton pump (for review see Stoeckenius and Bogomolni, 1982). The function of br is based on the characteristic absorption spectrum of the pigment = (X.,, 568 nm) and on its cyclic photoreaction. Both are closely associated with specific electrostatic and steric interactions between the retinyl polyene (chromophore I in Table 1) and its protein environment, to which it is linked via a protonated Schiff base bond. Acidification of purple membrane suspensions induces reversible transitions of br568 into two forms: one absorbing 605 nm (br605) with an apparent pk. of 2.9, and a second absorbing at 565 nm (br565) with pk, = 0.5 (Oesterhelt and Stoeckenius, 1971; Fischer and Oesterhelt, 1979). Later studies (Kimura et al., 1984; Chang et al., 1985) have shown that the H+ brs= br605 equilibrium is markedly affected by deionization of the membrane suspensions. Thus, when br-bound divalent cations are removed, the equilibrium shifts to br605 even in neutral solutions. On the other hand, the H+, ClbR.w5 - br_%5 equilibrium is unaffected by deionization. Considerable interest in br6o5 and br565, as clues to the structure and function of br, arose in relation to several points. (a) The absorption spectrum. Based on indirect evidence, it was suggested that the formation of br605 is associated with acid-induced changes in the interactions between the Schiff base and its counterion. It was proposed that such changes are induced either by direct titration (neutralization) of the intrinsic counterion (Fischer and Oesterhelt, 1979; Warshel and Ottolenghi, 1979; Smith and Mathies, 1985), or by a protein conformational change inducing an increased Schiff basecounterion separation (Fischer and Oesterhelt, 1979; Smith and Mathies, 1985; Szundi and Stoeckenius, 1987, 1988). The generation of br565 was attributed to protonation of a protein group in the vicinity of the f3-ionone ring (Warshel and Ottolenghi, 1979), or to binding of an extrinsic anion at the site responsible for the first transition (Fischer and Oesterhelt, 1979). More recently it has been suggested that low ph and/or high extrinsic ion Biophys. J. Biophysical Society Volume 56 December /89/12/1259/07 $ /89/12/1259/07 $

2 TABLE i Absorption maxima of arutifclal pigment at din Pigment Absorption maxima ph2.5 (br,!) ph -0.5 (br2) Spectral shifts AP,* I&2t nm nm cm-' cm-' N.>; Nt N,>, O II HN ON III X. N,N<, HO IV V Me e VI NO VII N N VIlI -1,-~ IX 1N NC NH NO ix , N xl Biophysical Journal Volume 56 November 1989

3 TABLE1 11.xN (continued) -CH O~ XII XIII XIV xv X xvi XIa xvii NH O XVIII OCH, xix xx CN~~~.,,H xxi N XXII *Difference in energy between absorption maxima at ph 7 and ph -2.5 (first acid transition). tdifference in energy between absorption maxima at ph -0.5 and ph -2.5 (second acid transition). The absorption band changes after -60 s to a band at 450 nm. See also Gartner and Oesterhelt, Albeck et al. Absorption Maxima of Bacteriorhodopsin 1261 Absorption Maxima of Bacteriorhodopsin 1261

4 concentrations induce the formation of br565 by restoring a Schiff base environment similar to that of the native brs68 chromophore (Szundi and Stoeckenius, 1987, 1988). (b) The photocycles. Both br6os (Mowery et al., 1979; Kobayashi et al., 1983; Dupuis et al., 1985; Chang et al., 1985; Ohtani et al., 1986) and br565 (Mowery et al., 1979) exhibit photocycles which differ from that of br568. Especially relevant is the observation that brw5 lacks the characteristic M412 photointermediate of br568 and does not pump protons. (c) Cation binding in br. Binding sites and binding mechanisms and their relationships with the membrane surface potential, with the protein conformation, and with photodeprotonation of the Schiff base, are all closely associated with the br568 br6o5 equilibrium (Chang et al., 1986; Dunach et al., 1987; Concoran et al., 1987; Szundi and Stoeckenius, 1987, 1988; Ariki et al. 1987). (d) The direction of the exposure of the Schiff base in respect to the membrane surface (Druckmann et al., 1985). (e) The kinetic mechanism of light - dark (all-trans 1 3-cis) adaptation (Warshel and Ottolenghi, 1979). In spite of such extensive activity, the molecular changes in the protein and in the protein-retinal interactions responsible for the generation of br6o5 and br565 are still unclear. In the present work we approach the problem by applying artificial br pigments in which the native retinal has been replaced by synthetic analogues (for reviews see Derguini and Nakanishi, 1986; Sheves et al., 1987). The method has substantially contributed to the understanding of the color and photocycle of br. It now bears on the identification of specific regions in the polyene-binding site, associated with the formation of the acid forms of bacteriorhodopsin. MATERIALS AND METHODS The synthesis of the retinal analogues has been described previously (for review see Ottolenghi and Sheves, in press. For pigment XVIII see Gartner et al., 1983; for XIX see Gartner and Oesterhelt, 1988; for XX-XXII see Derguini et al., 1983; Sheves and Friedman, 1986). The artificial pigments were prepared by reconstituting the apomembrane suspended in water with the retinal analogues at 25 C (Tokunaga and Ebrey, 1978). The acidic forms of the artificial pigments were obtained by mixing the pigments with appropriate buffers. Absorption spectra were measured using a model 8450A diode array spectrophotometer (Hewlett-Packard Co., Palo Alto, CA). RESULTS AND DISCUSSION The effects of acid on the absorption spectra of a variety of artificial light-adapted br pigments are summarized in Table 1. The effects may be classified according to three main categories: (a) pigments II-VI which, analogously to br568 (I), exhibit two spectral transitions: a red shift with an apparent pk. -2.5, followed by a blue shift with pk (b) pigments (VII-XIX), which exhibit only the first transition, and (c) pigments whose spectra are unaffected by acid down to ph -0.5 (XX-XXII). It is worthwhile noting that in most cases of categories a and b, the magnitude of the spectral shift associated with the first transition (Av, = 1,100 ± 350 cm-') is reasonably close to that characterizing the br568 br6o5 interconversion in the native pigment (Av = 1,015 cm-'). Somehow higher positive or negative deviations from the latter value are observed in the cases of III, IV, and XVI. We thus conclude that with class a and b pigments we observe the generation of acid species (br',, bri,, etc.) which are analogous to brw (bros). The same applies in the case of class a to the parameter AP2. The latter measures the blue shift (1,200 ± 550 cm-') associated with the conversion of the first acid species to the highly acidic forms (br'i, br',,, etc.) which are thus assumed to be analogous to brl (br565). We first consider the formation of the br! species which bear on the structure of br605. These are observed even when the polyene chain is seriously perturbed, including the extreme case of VII in which only one C=C bond is present. This implies that the respective spectral shift must be induced by changes in protein-retinal interactions in the vicinity of the C13-NH+ region of the chromophore. A plausible specific mechanism, in keeping with previous suggestions (Warshel and Ottolenghi, 1979; Fischer and Oesterhelt, 1979) and with current models for the spectrum of br (Lugtenburg et al., 1986; Spudich et al., 1986; Ottolenghi and Sheves, in press), may be based on the direct titration of a negative ion (counterion) in the vicinity of the protonated Schiff base. However, alternative (indirect) mechanisms associated with acidinduced changes in the conformation of the protein in the vicinity of Schiff base are also feasible. For example, an increased nitrogen counterion separation (Fischer and Oesterhelt, 1979; Smith and Mathies, 1985; Szundi and Stoeckenius, 1987, 1988), or changes in H-bonding to the nitrogen of residual water molecules or of other protein residues (Warshel and Barboy, 1982; Baasov and Sheves, 1986), may both be associated with a red spectral shift. The only exceptions in respect to br! formation are the class c, cyaninelike pigments. Their insensitivity to acidification is readily accounted for by the basic insensitivity of the spectra of their parent (symmetric) cyanine dyes to nonconjugated charges, as well as to H-bonding to the NH' moiety (Sheves and Friedman, 1986) Evidence supporting the conclusions that the generation of br605 is due to a structural perturbation in the vicinity of the Schiff base linkage may also be derived from 13C NMR studies of br (Harbison et al., 1985) and of model protonated Schiff bases in solution (Albeck, A., 1262 Biophysical Journal Volume 56 November Biophysical Journal Volume 56 November 1989

5 H. Gottlieb, and M. Sheves, manuscript in preparation). A significant perturbation is observed in the `3C5 chemical shift in br (144.5 ppm), which is shifted by 12 ppm downfield, relative to a model system (131 ppm). This dramatic effect on the chemical shift is partially due to the planar, s-trans, ring-chain conformation which prevails in the pigment. However, studies of model systems (Albeck, A., H. Gottlieb, and M. Sheves, manuscript in preparation) indicated that the 13C5 chemical shift is also sensitive to changes in the Schiff base environment. Weakening of hydrogen bonding, which red shifts the spectrum, also affects the C5 chemical shift. The effect is especially pronounced in the case of the s-trans ring-chain conformation. Model systems that mimicked the weakly H-bonded Schiff base of br, exhibited a downfield shift of -5 ppm in the C5 chemical shift. Thus, assuming that the spectral shift associated with br6o5 is due to perturbation of the Schiff base environment, such studies would predict an additional shift of 2-4 ppm for C5 in br6o5. This is in keeping with the -5 ppm value measured in br605 by solid state 13C NMR (de Groot et al., 1988). We now consider the second acid-induced transition, from br! (analogous to br605) to br3 (analogous to br565), observed in class a pigments in which the basic polyene system of br568 is only slightly modified. Major changes in the ring region such as replacement by an aromatic ring (VII-XII), shortening the polyene chain (VII-XI) or of the polyene sequence (XV-XVI), and bulky substitution at the C4 position (XIII-XIV) lead to elimination of the second transitions. Molecular models accounting for these observations should be considered in light of the present approach to the spectrum of bacteriorhodopsin. The latter is quantitatively described by the "Opsin Shift" (OS) (Nakanishi et al., 1980), which measures the energy difference between the absorption maximum of a model protonated Schiff base in methanol solution (440 nm) and that of the pigment (568 nm). Accumulated evidence is now available (Baasov and Sheves, 1985; Harbison et al., 1985; Lugtenburg et al., 1986; Spudich et al., 1986) indicating that the observed value of OS in bacteriorhodopsin (-5,000 cm-') is the result of three major protein effects with contributions denoted by OS,, OSp, and OSd. The first term accounts for more than half of the observed shift (i.e., OS 3,000 cm-' ) and is attributed to electrostatic interactions involving the Schiff base nitrogen such as nitrogen-counterion separation, H-bonds, local dielectric factors, etc. The second term is attributed to a protein-induced planar ring-chain configuration (OSp 1,300 cm-') and the third to the presence of a protein ion pair (or a dipole) in the vicinity of the ring - (OSd 700 cm-'). On the basis of this model for the binding site, two approaches, accounting for the - br605 br565 transition, may be advanced. The first is based on perturbation of the environment of the ring moiety. This can be carried out in various ways; e.g., Warshel and Ottolenghi (1979) suggested neutralization of a ring charge (ion pair) via direct titration. However, it has been suggested that the br605 br565 transition is induced by extrinsic ions rather than by H+ (Fischer and Oesterhelt, 1979; Kimura et al., 1984). In such a case, it will be more feasible to suggest that the effect is due to an increase in the polyene-ion pair distance due to an (indirect) acidinduced conformational change. Alternatively, it is possible that such a change induces a relaxation of the ring-chain conformation, from planar (s-trans) to distorted (s-cis), which will also result in a blue shift. Either of such alternatives will account for the lack of the second acid transition in artificial pigments bearing a seriously perturbated ring region. Moreover, it is expected that the value of Av2 for class a pigment will be comparable to Sd or to OSp, as it is actually confirmed by the data of Table I. Because, AV2 is substantially smaller than OSp + OSd, the possibility that both acid-induced effects take place may be excluded. The second approach interprets the spectrum of br565 in terms of changes induced in the vicinity of the Schiff base linkage, restoring an environment similar to that of br568. Thus, the previous explanation, which excludes changes in the Schiff base environment as the source for the blue shift in br565, confronts difficulties with resonance-raman measurements of acidified forms of bacteriorhodopsin (Smith and Mathies, 1985), indicating different C=N stretching frequencies for br605 (1,630 cm-') and br565 (1,637 cm-'). This observation might reflect different isomer ratios (13-cis and all-trans) in both forms (Smith and Mathies, 1985) and/or a different Schiff base environment. For example, weak hydrogen bonding to the positively charged nitrogen with its counterion or with protein dipoles (or residual water) shifts the C=N frequency to a lower energy (Baasov et al., 1987; Rodman Gilson et al., 1988) and induces a red shift in the absorption spectrum. Thus, the change in the C=N stretching frequency in br565 relative to br605 might reflect stronger hydrogen bonding to the Schiff base linkage in the former, causing a blue shift in the spectrum. Obviously, in this case, the question arises as to why artificial pigments bearing a seriously perturbed ring region lack the second acid transition. A possible explanation implies an intimate coupling between the (tight) chromophore-protein conformation in the vicinity of the ring region and the change taking place in the Schiff base linkage environment in the second acid transition. Modification of the ring region in pigments VII-XVI perturbs the chromophore-protein interaction, preventing the change in the Schiff base environment in low ph. The mechanisms suggested to account for the second Albeck et al. Absorption Maxima of Bacteriorhodopsin Albeck et al. Absorption Maxima of Bacteriorhodopsin 1 263

6 acid transition should also be considered in view of pigments XVII-XIX which, in spite of their basically unmodified polyene sequence, do not exhibit the transition to br?. We recall that such pigments differ from all others in Table 1 in carrying a 1 3-cis, rather than all-trans, chromophore (Gartner et al. 1983; Albeck et al. 1986). However, it appears that the 13-cis isomer of br has a planar ring-chain conformation analogous to that of the all-trans isomer (Harbison et al., 1985). Thus, it is tempting to suggest that the 13-cis isomers lack the dipole-ring interactions which characterize analogous alltrans pigments and, therefore, do not exhibit transition to br?-like species. This interpretation will also qualitatively account for the decreased opsin shift (500-1,500 cm-' ) of the above 1 3-cis isomers in respect to analogous all-trans chromophores, such as I-IV. An alternative explanation for the observation that the 1 3-cis analogues lack the br? transition may be associated with the finding that the br br565 transition is coupled with a 13-cis all-trans isomerization of the chromophore (Mowery et al., 1979, Smith and Mathies, 1985). Because chromophores that are a priori 1 3-cis (in both br568 and br605) are unable to undergo such an isomerization, they will not generate the br? species. CONCLUSIONS The application of artificial br pigments provides direct experimental evidence indicating that the acid (or deionized) br605 species is associated with changes in proteinpolyene interactions in the vicinity of the Schiff base. The data are in keeping with the direct titration of a negative or dipolar protein group. Alternatively, titration of a protein group outside of the retinal binding site is also feasible. In such a case it is implied that the titration induces changes in the protein conformation that affect protein-retinal interactions in the vicinity of the Schiff base linkage. No definite conclusions can be derived in relation to the structure of the br565 species. Thus, our data cannot discriminate between a mechanism based on changes in the protein environment around the f3-ionone ring and one invoking changes in the vicinity of the Schiff base. If the latter applies, it is implied that such changes are conditioned by an intact chromophore structure in the ring region. We thank Prof. W. Stoeckenius for valuable comments. We acknowledge support from the United States-Israel Binational Science Foundation and the Fund for Basic Research administered by Israel National Academy of Sciences and Humanities, and support from the Kimmelman Center of biomolecular structure and assembly. REFERENCES Albeck, A., N. Friedman, M. Sheves, and M. Ottolenghi Role of retinal isomerizations and rotations in the photocycle of bacteriorhodopsin. J. Am. Chem. Soc. 108: Ariki, M., D. Madge, and J. K. Lamb Metal ion binding sites of bacteriorhodopsin. J. Biol. Chem. 262: Baasov, T., and M. Sheves Model compounds for the study of spectroscopic properties of visual pigments and bacteriorhodopsin. J. Am. Chem. Soc. 107: Baasov, T., and M. Sheves Alternation of pka of the Schiff base of bacteriorhodopsin. A study with model compounds. Biochemistry. 25: Baasov, T., N. Friedman, and M. Sheves Factors affecting the C=N stretching frequency of protonated retinal Schiff base. A model study for bacteriorhodopsin and visual pigments. Biochemistry. 26: Chang, G. H., J. G. Chen, R. Govindjee, and T. Ebrey Cation binding by bacteriorhodopsin. Proc. Nati. Acad. Sci. USA. 82: Chang, C. H., R. J. Jonas, S. Melchiou, and R. Govindjee Mechanism and role of divalent cation binding of bacteriorhodopsin. Biophys. J. 49: Concoran, C. T., K. Z. Ismail, and M.A. El-Sayed Evidence for the involvement of more than one metal cation in the Schiff base deprotonation process during the photocycle of bacteriorhodopsin. Proc. Natl. Acad. Sci. USA. 84: de Groot, H., S. Smith, J. Curtin, E. van der Berg, C. Winkel, J. Lugtenburg, J. Herzfeld, and R. Griffin High resolution solid state NMR of blue membrane. Biophys. J. 53:381a (Abstr.) Derguini, F., and K. Nakanishi Synthetic rhodopsin analogs. Photobiochem. Photobiophys. 13: Derguini, F., C. Caldwell, M. Motto, V. Balogh-Nair, and K. Nakanishi Bacteriorhodopsins containing cyanine dye chromophores. Support for the external point-charge model. J. Am. Chem. Soc. 105: Druckmann, S., M. Ottolenghi, A. Pande, J. Pande, and R. Callender Acid-base equilibrium of the Schiff base in bacteriorhodopsin. Biochemistry. 21: Dunach, M., M. Seigneuret, J.-L. Rigand, and E. Padros Characterization of the cation binding sites of the purple membrane. Electron spin resonance and flash photolysis studies. Bichemistry. 26: Dupuis, P., T. C. Corcoran, and M. A. El-Sayed Importance of bound divalent cations to the tyrosine deprotonation during the photocycle of bacteriorhodopsin. Proc. Natl. Acad. Sci. USA. 82: Fischer, U., and D. Oesterhelt Changes in the protonation state of bacterio-opsin during reconstitution of bacteriorhodopsin. Biophys. J. 28: Gartner, W., and D. Oesterhelt Methoxy retinals in bacteriorhodopsin. Eur. J. Biochem. 176: Gartner, W., S. Towner, H. Hopf, and D. Oesterhelt Removal of methyl groups from retinal controls the activity of bacteriorhodopsin. Biochemistry. 22: Harbison, G. S., S. 0. Smith, J. A. Pardoen, J. M. Courtin, J. Lugtenburg, J. Herzfeld, R. A. Mathies, and R. G. Griffin Solid state '3CNMR detection of perturbed 6-s-trans chromophore in bacteriorhodopsin. Biochemistry. 24: Biophysical Journal Volume 56 November 1989

7 Kimura, Y., A. Ikegami, and W. Stoeckenius Salt and phdependent changes of the purple membrane absorption spectrum. Photochem. Photobiol. 40: Kobayashi, T., H. Ohtani, J. Iway, A. Ikegami, and H. Uchiki Effect of ph on the photoreaction cycles of bacteriorhodopsin. FEBS (Fed. Eur. Biochem. Soc.) Lett. 162: Lugtenburg, J., M. Muradin-Szweykowska, C. Heeremans, and J. A. Pardoen Mechanism for the opsin shift of retinal's absorption in bacteriorhodopsin. J. Am. Chem. Soc. 108: Mowery, P. C., R. H. Lozier, Q. Chae, Y.-W. Tseng, M. Taylor, and W. Stoeckenius Effect of ph on the absorption spectra and photoreactions of bacteriorhodopsin. Biochemistry. 18: Nakanishi, K., V. Balogh-Nair, M. Arnaboldi, K. Tsujimoto, and B. Honig An external point-charge model for bacteriorhodopsin to account for its purple color. J. Am. Chem. Soc. 102: Oesterhelt, D., and W. Stoeckenius Rhodopsin-like protein from the purple membrane of Halobacterium halobium. Nat. New Biol. 233: Ohtani, H., T. Kobayashi, and J.-I. Iwai Picosecond and nanosecond spectroscopies of the photochemical cycles of acidified bacteriorhodopsin. Biochemistry. 25: Ottolenghi, M., and M. Sheves Synthetic retinals as probes for the binding site and photoreactions in rhodopsins. J. Membr. Biol. In press. Rodman Gilson, H., B. Honig, A. Croteau, G. Zarrilli, and K. Nakanishi Analysis of the factors that influence the C=-N stretching frequency of polyene Schiff bases. Biophys. J. 53: Sheves, M., and N. Friedman Influence of external negative charges on the absorption maxima of symmetrical cyanines. A study with model compounds and artificial bacteriorhodopsin pigments. Ang. Chem. Int. Ed. Engi. 25: Sheves, M., A. Albeck, T. Baasov, N. Friedman, and M. Ottolenghi The binding site and molecular changes in the photocycle of bacteriorhodopsin. Studies with synthetic retinal analogs. In Retinal Proteins. VNU Science Press, Utrecht, The Netherlands Smith, S., and R. Mathies Resonance Raman spectra of the acidified and deionized forms of bacteriorhodopsin. Biophys. J. 47: Spudich, J. L., D. A. McCain, K. Nakanishi, M. Okabe, N. Shimizu, H. Rodman, B. Honig, and R. A. Bogomolni Chromophoreprotein interaction in bacterial sensory rhodopsin and bacteriorhodopsin. Biophys. J. 49: Stoeckenius, W., and R. A. Bogomolni Bacteriorhodopsin and related pigments of halobacterium. Annu. Rev. Biochem. 51: Szundi, I., and W. Stoeckenius Effect of liquid surface charges on the purple to blue transition of bacteriorhodopsin. Proc. Nall. Acad. Sci. USA. 84: Szundi, I., and W. Stoeckenius Purple to blue transition of bacteriorhodopsin in a neutral liquid environment. Biophys. J. 54: Tokunaga, F., and T. Ebrey The blue membrane: the 3-dehydroretinal-based artificial pigment of the purple membrane. Biochemistry. 17: Warshel, A., and N. Barboy Energy storage and reaction pathways in the first step of the vision process. J. Am. Chem. Soc. 104: Warshel, A., and M. Ottolenghi Kinetic and spectroscopic effects of protein-chromophore electrostatic interactions in bacteriorhodopsin. Photochem. Photobiol. 30: Albeck et al. Absorption Maxima of Bacteriorhodopsin 1265